XSF Discussion - 2018-07-31

mind, even CLIENT: <foo/> SERVER:<bar/> can easily be parsed as XML

yeah, but I do believe we should discourage that and encourge one <example/> per stream element or stream element group

I'm not entirely sure about that, I think having a single logical example in a single example block makes reading easier sometimes.

Yes. It's better to have an exchange in one example than in two or four

+1 Kev, Ge0rG

https://slack.engineering/reducing-slacks-memory-footprint-4480fec7e8eb

it has a nice infographic on how slack handles notifications

If we care about an example being valid XML without annotation (and I'm not sure that we do, but maybe), <!-- Client --> <!-- Server --> work.

<https://cdn-images-1.medium.com/max/2000/0*PV_09olld6K1l8jQ.> for those who don’t like javascript (yes, the dot is part of the URL)

And if we can get the (whatever component does the styling) to special-color "client:" and "server:" at the highest level, it'd be great.

Kev, the text in front of the elements is valid XML, too ;-)

Well, yes, that's true.

jonasw: are you saying that any text is valid XML?

Ge0rG, no

But only if you wrap it all.

Kev, you have to do that anyways when you have multiple elements

Kev: valid XML, sure. Valid XMPP, no :)

phew. I was frightened for a moment.

jonasw: Yes, true.

intosi, comments wouldn’t be valid XMPP either

intosi: Sure, but ... I don't think we're going to validate XMPP in the examples.

not that strictly at least. I do have a prototype of a thing which validates the examples against the respective XML schemas

Kev: but I was just going to copy-paste all XEP examples into my client's XML console!

Ge0rG: I wouldn't try to stop you :)

jonasw: That Slack article is very much the direction I've been pushing with XMPP, too.

Having servers know about unread counts and understand notifications.

Let's make XMPP a Slack clone!

Kev: but the meta data! It's getting even more! And worse!

Ge0rG: Or, let's use good ideas even if other people have them too (after us, in this case).

> Ge0rG> Yes. It's better to have an exchange in one example than in two or four

Not sure, the examples in RFC 6120 § 9 a very readable without using one big example block

flow: btw... did you want to offer me your help in bringing smack4 into yaxim, again? :)

Ge0rG, No, I didn't want to offer my help, I did offer my help

and that offer still stands

flow: the markup in https://xmpp.org/rfcs/rfc6120.html#examples is rather minimal, the one in https://xmpp.org/extensions/xep-0409.html#activation is rather distracting imho

Ge0rG, yes, but that is a layout issue, maybe we need to invent an additinal <example/> thingy for sequences

> This application can consume between ~130MB (p10) and ~960MB (p99) of memory depending on team activity and UI state.

damn

my point is that the thing that example encloses should be just XMPP XML, and not surrounding annotations. Using XML comments for annotations doesn't appeal to me too, because it may give the impression that XMPP allows them

hm

IIRC Link Mauve (?) worked on validating examples with the provided schema. I'd really like to see that happen, as it helps finding issues in the examples *and* the schema

we can do that already

source: I have code which does it

and one precondition would probably be that the, possibly new to be created, <examples-foo/> thing is as pure as possible

jonasw, is it part of your xeplint?

yes

or at least of an earlier version

I think the current one doesn’t do it yet

that's great, how many false negatives to you get?

the issue is that many included schemas aren’t available

so validation breaks very easily

"aren't available"?

404

are those schemas not included in the XEP?

the ones defined by the XEP are, but sometimes they refer to other XEP’s schemas

jonasw, ahh, but then it's not really a 404 if the other XEP is availble, isn't it? :)

flow, it is, because the schema is included via an HTTP url

guessing from which XEP the included schema originates is impossible without scanning over all XEPs

Kev, and do you have a proposol on how to get anything nearly as complex as that graph standardized and configurable?

Not yet. My first interest is solving the unread problem, and then moving on to notifications.

FWIW, a lot of it revolves around attaching references to messages, I think.

I think if we craft references sensibly, we can at least make mentions and things easier for a server to process.

Kev, is there any hope that my feedback gets incorporated into References?

I've not touched References in a while, sorry. If you gave feedback and I've not addressed it, it should still be in my XSF TODO folder.

Edwin's got the lead of References updates at the moment.

2018-03-12 is the date of my email, if that helps

Although I keep giving him other stuff to do :)

Is the "unread problem" blocking bind2?

No, I don't think so.

do we have an agenda?

I'd like to add a proto XEP

Ge0rG, We do; I've just not quite sent it yet. What's the ProtoXEP you wanted?

Dave Cridland: I need to submit it yet.

synopsis?

Ge0rG, Hurry!

Dave Cridland, do you have file sharing notifications on your list?

jonasw, "fsn" and #681.

jonasw: XEP-0050: Remove the status attribute from the request #681 https://github.com/xsf/xeps/pull/681

> XEP-xxxx: Schrödinger's Chat > This protocol extension for XEP-0045 Multi User Chat allows clients to check whether they are still joined to a chatroom.

Dave Cridland, #682

Dave Cridland: XEP-0065, XEP-0363: Document exposing the service on the user’s domain #682 https://github.com/xsf/xeps/pull/682

I love the name.

It's been the name of the XMPP off-topic MUC for a while now.

Dave Cridland, the vote on #672 is missing from your spreadsheet of doom

Dave Cridland: XEP-0059: Add 'exact(-index)' attribute to RSM's <count/> and <first/> #672 https://github.com/xsf/xeps/pull/672

(you probably know that, but I thought I’d mention it anyways)

Ted Sterr actually put the voting stuff into a CSV and sent it to me to paste in. I need to do that (maybe Real Soon Now)

awesome

can I subscribe to that?

would be useful to keep the pull requests clean

jonasw, It's public; let me dig out the link.

jonasw, https://docs.google.com/spreadsheets/d/1AZ-Sna6OiRG--b-mJMKv3XXfrn3Nehm0kAtlyJvImL0/edit?usp=sharing

I know your spreadsheet (I have it bookmarked, even), but the CSV service from Ted would be good too :)

also, was #664 discussed?

jonasw: XEP-0045: Add implementation note about {jabber:x:conference}x payload #664 https://github.com/xsf/xeps/pull/664

jonasw, Oh. I don't think it's a *service* as such...

oh, it was

totally missed that all were +1 except for the typo :)

jonasw, curl https://docs.google.com/spreadsheets/d/1AZ-Sna6OiRG--b-mJMKv3XXfrn3Nehm0kAtlyJvImL0/export?format=csv ?

MattJ, no, I don’t need CSV specifically. I had the impression that Ted had some updated data which isn’t in that document yet

jonasw, I'll put that in now. Sadly doesn't cut/paste.

jonasw, If you want edit rights to the sheet, though, just ask.

if I were able to follow the council meetings closely enough to do that, I’d be taking minutes and/or not need that sheet at all, unfortunately :(

Ge0rG, How imminent is that ProtoXEP? ✏

Dave Cridland: it's 75% ready.

Dave Cridland: if you put it on the agenda, I'll get it to the editor within the next 2h

the editor can’t promise to manage it to get it on the website and emailed today unless there is a PR before 18:30 CEST ✏

but I’ll do my best

jonasw: 18:30, okay.

Ge0rG, I can't send the agenda out with a protoxep on it that's not been announced. Don't rush it - we'll do it next week if not this.

Dave Cridland: I don't believe into those delays, and I'm going on holiday in three weeks.

Well, I'll hold off as long as possible.

Dave Cridland, what’s your limit for delaying the agenda?

Well, I (try to) send it out 24 hours in advance, which was 26 minutes ago. :-)

But in practise, Kev only starts complaining if it's not out by tomorrow morning.

And Kev's the only person who ever complains.

I didn't say that.

Hinted, very strongly, but didn't actually say...

Which might, disappointingly, mean Kev's the only person who puts aside time to review the agendums before each meeting.

I try to take aside the time. I'm just too polite to complain.

> Copyright: © 1999 – 2017 XMPP Standards Foundation. SEE LEGAL NOTICES. we should fix the year

Ge0rG, Easier to fix the copyright statement.

Dave Cridland: do we have a process for *that*?

jonasw: https://github.com/xsf/xeps/pull/687

wee

Rendered version: https://op-co.de/tmp/muc-selfping.html

build in progress

Ge0rG, for optimum speed, watch https://hub.docker.com/r/xmppxsf/xeps/builds/ and tell me when the newest goes green

There is a huge red "Cancel" button.

I. Must. Resist.

Ge0rG, which won’t do anything unless you’re logged in and have permissions

funnily enough, the response of the cancel button is exactly the same if you’re logged in or not.

you need to reload to find out whether it actually cancelled something

Ge0rG, Where did you see that copyright statement? The one in xep.xsl looks to be 2018.

Dave Cridland, look at the date of the last commit touching that file ;-)

Ah, 5 minutes ago...

(TL;DR: I patched it just five minutes ago) ✏

> Building

Ge0rG, Still building?

Dave Cridland: still building.

Dave Cridland: Jonas said, it might take 90mins

Madness

Dave Cridland: ETA is some time between 30 and 55mins from now. You could also wait until tomorrow morning and tell Kev to complain to me about it.

Well, got the agenda item all written up ready, just got to hit sent when it's done.

That's great preparation.

For once.

Ge0rG, Dave Cridland, check your (standards@) inboxes, the ProtoXEP is now at https://xmpp.org/extensions/inbox/muc-selfping.html

Agenda launched!

impact in 3 ... 2 ... 1 ...

Damn, my mobile email client has reset itself.

SamWhited: "MUC Self Ping" is way too unexciting

Good. I like unexciting.

(Schrödinger's Chat is very amusing though; maybe that could be an april fools XEP or something)

MattJ: What was the term you pointed out in the ActivityPub spec?

SamWhited: Boring specs are out

> cool linked data

Cool!

Standards are slipping

If it doesn't fit in 140 characters then it's too long and complicated!!11

-xep serious

Zash: Bookmarks 2 (This Time it's Serious) (Standards Track, Experimental, 2018-07-22) See: https://xmpp.org/extensions/xep-0402.html

QED

> I love the name. -- Dave Cridland

That one tells you what it does in a clear way before making a joke, at least.

So we can settle on "MUC Self Ping (Schrödinger's Chat)"?

I'd rather prefer it to be the other way round

Schrödinger's chat (Or, how I learned to stop worrying and self-ping in MUCs)

This one weird trick...

Zash [18:56]: > Schrödinger's chat (Or, how I learned to stop worrying and self-ping in MUCs) 👍

Funn, just a few days ago I exectly implemented self ping in a bot I wrote.

martin: did you do everything right?

Hope so. I send ping to muc participant and check for result IQ with the correct ID. At least it works. Have not yet looked at your proposal in detail.

Ge0rG: uh, your proposal says server reply is type 'get'. I always got 'result' or 'cancel'.

I looked at 0199 so I did it this way.

martin: sorry, copy & paste error in the example

martin: no, wait. It's not the response, it's the reflection

I see.